N.sup.2 -((S)- 1-ethoxycarbonyl-3-phenylpropyl)-N.sup.6 -trifluoroacetyl-L-lysyl-L-proline can be prepared by a method described in Japanese Unexamined Patent Publication No. 36297/1986, and the like. As a method for isolation and purification of thus obtained N.sup.2 -((S)-1-ethoxycarbonyl-3-phenylpropyl)-N.sup.6 -trifluoroacetyl-L-lysyl-L-proline, crystallizing method is extremely easy. In Japanese Unexamined Patent Publication No. 36297/1986, there is described a crystallizing method using methyl tert-butyl ether/cyclohexane (650 ml/240 ml). Also, in Journal of Organic Chemistry, 53, p836-843 (1988), there is disclosed a crystallizing method using methyl tert-butyl ether/cyclohexane or methyl tert-butyl ether/hexanes (C.sub.6 H.sub.14) (225 ml/85 ml cyclohexane). However, as to the methods using these solvents as crystallizing solvents, there are problems that the solvents are highly hazardous and that the crystal properties of the resulting N.sup.2 -((S)-1-ethoxycarbonyl-3-phenylpropyl)-N.sup.6 -trifluoroacetyl-L-lysyl-L-proline is unfavorable. Therefore these methods are not necessarily preferable. Due to the unfavorable crystal properties, solid-liquid separation takes long period and mother liquor cannot be sufficiently replaced when washing cake. Further, once obtained crystal easily melts or becomes oily (syrupy). So, during solid liquid separating step to drying step, there arise following problems. That is, the molten or oily state of some or almost of crystal results in unsuccessful separation and collection thereof. Also, the extreme difficulty of the removal of solvents from the obtained oily matter or molten matter, requires long period for drying step or results in stopping of drying step halfway.
Preliminary research of the present inventors for crystallizing solvents other than methyl tert-butyl ether/cyclohexane or methyl tert-butyl ether/hexanes (C.sub.6 H.sub.14) has shown problems that N.sup.2 -((S)-1-ethoxycarbonyl-3-phenylpropyl)-N.sup.6 -trifluoroacetyl-L-lysyl-L-proline is not crystallized or is obtained only in an oily state, or even if gelled matter obtained, collection by filtration cannot be done due to unsuccessful crystallization. Thus it has been proved that the crystallization is an extremely difficult technical subject.
An object to be solved by the present invention is to find out solvents which enable the crystallization of N.sup.2 -((S)-1-ethoxycarbonyl-3-phenylpropyl)-N.sup.6 -trifluoroacetyl-L-lysyl-L-proline with favorable crystal properties, instead of methyl tert-butyl ether/cyclohexane or methyl tert-butyl ether/hexanes (C.sub.6 H.sub.14).